THE EFFECTS OF NOISE FILTERS ON SEGMENTATION BASED SEEDED REGION GROWING

Abstract

Image segmentation is a process of grouping pixels to make parts of objects into distinct image areas using their texture, edge, color properties. The segmentation process plays an important role in the analysis of images and in image processing. One of the techniques developed for segmentation is SRG (Seeded Region Growing). The noise generated during the acquisition of images affects the segmentation success negatively. Filters used to eliminate noise reduce it, but the effect of filtering on the segmentation success is not fully known. In this study, the effects of noise and filters on the SRG algorithm are investigated. For this purpose, various noises were added to Weizmann database images at different levels. Later, filters were applied to noisy images. Finally, F-Score values were obtained from the images segmented by the SRG algorithm and compared with the values of the original images.